Locomotives in mid train

[SD70Dude]

And since this thread is about DP/Locotrol, here's the brief Transport Canada summary of another, much more recent incident involving a switching move gone wrong.  A poster on another forum who is from the area stated that it got up to over 80 mph.

Accident/Incident Type:UNCONTROLLED MOVEMENT OF
R/S
INCIDENT
2020-03-10 16:00 Classification: 5
Province: Nearest Town/City:
Subdivision Owner:
Subdivision Name:
SASKATCHEWAN DANA
CN - CANADIAN NATIONAL RAILWAY CO.
Aberdeen Subdivision Mileage: 24.60
Location:
Details:
DG Cars Involved:
Train Operator:
Yard Name:
0
CN - CANADIAN NATIONAL RAILWAY CO.
DG Released: 0
Spur Name: Spur Mileage:
Injuries:
Fatal: 0 Serious: 0 Minor: 0
Occurrence Summary:
CN assignment L52442-10 operating light engine (CN 3893, CN 3181, UP 7920) set out unit UP 7920 on the Main track at mile 24.6
Aberdeen Sub and reversed up the Totzke Spur to lift their potash loads. At the same time unit UP 7920 moved Eastward,
uncontrolled. The crew received an alert from the UP 7920 and stopped the unit at mile 19.5 Aberdeen Sub. after going over 5
crossings. No conflicting movements. No derailment. No injuries.

[SD70Dude]

I don't think we have a limitation on the number of locomotives that can be coupled together.  I've seen a few power moves of well over 10 units.  As for pulling power, CN's maximums are 24 powered DC axles (30 on certain subdivisions) per consist, this drops to 18 if you have even one AC unit.  Dynamic braking is restricted to 18 axles per consist, 12 if you have an AC unit. 

It is possible to cut out (disable) the DB on locomotives, and some newer units will allow you to place them in 'DB only' mode, which would be used when that unit is not to be used for pulling power in order to comply with the much-hated throttle restrictions. 

[BaltACD]

Lithonia Operator

What's the largest total number of units you've ever seen on a train?

 

The Chessie System rule permitted a maximum of 6 units to be coupled together.

CSX has moved the maximum number into their "Train Handling Rules" and out of the Book of Rule and the Employee Timetable - I believe the allowed maximum in a single engine consist is 12 units.  The number of units in the engine consist that can actually be on line is further restricted by the number of powered axles.  When counting powered axles AC locomotives count as 9.

 

[MMLDelete]

What's the largest total number of units you've ever seen on a train?

[jeffhergert]

There can be one lead consist and up to 4 remote consists.

Jeff

[BaltACD]

steve-in-kville

So, if two loco's are up front, the lead controls the second by wired connection. But if its mid-train or on the end, its radio controlled?

Locomotives in a coupled engine consist are wired to each other with the 27 pin MU Cable, as well as the coupling of several pneumatic hoses required for proper operations.

Locomotives positioned elsewhere in the train are controlled by radio commands initiated by the engineer on the engine consist that is leading the train.  Those DPU consists may be one or more locomotives and they will be coupled with a MU Cable in their own engine consist.

[steve-in-kville]

So, if two loco's are up front, the lead controls the second by wired connection. But if its mid-train or on the end, its radio controlled?

[SD70Dude]

Overmod

While we are on the subject of ignorance of Locotrol, what was the Canadian case involving the two midtrain helpers left configured to 'pull the other way'?  You provided the reference in a different thread but search won't find it now, and I can't tickle Google with the right search params to cough it up.

Apologies for the late reply, it's this one.  The infamous Cheakamus Canyon caustic soda spill:

https://tsb.gc.ca/eng/rapports-reports/rail/2005/r05v0141/r05v0141.html

 

[MMLDelete]

tree68

 

 

Semper Vaporo

One thing that trains have always been ahead of the technological curve is in automatic steering!

 

 

And sometimes that doesn't work, either...

 

That reminds me of a story I heard about the legendary stock car driver Curtis Turner.

He crashed heavily into the wall at Daytona. Once he climbed out of the crumpled car, a reporter asked, "Curtis, what happened?"

Turner said, "I reckon I got a little behind on my steering."

 

[Electroliner 1935]

BaltACD

After all they had spent million$ with 'Union Switch & Swindle' to have the system created and installed.

Nice to hear that the old U.S. & S. nickname is still in use. Learned that back in 1956.

[BaltACD]

jeffhergert

But when you've spent millions of dollars on something, it's all a success even if it isn't.  And EMS isn't the first time that's happened on the railroad.

I've been told the Trip Optimizer algorithms are the same ones used for PTC to calculate braking distances.  

Jeff

When I first went to Jacksonville in 1990 and CADS was the 'new toy' of Dispatching, management for the most part wanted CTC territories to be run using the 'automatic' tools that were apart of CADS.  After all they had spent million$ with 'Union Switch & Swindle' to have the system created and installed.

That operating dictum only lasted for a year or so - the situations that 'automatic' created ended up taking hours and days for the territories to recover from the situations the system had created.

My personal observations were that neither the company or US&S devoted any resources to diagnose the problems and create solutions to the issues that were being observed.

With CADS being in its 3rd or 4th iteration, total automatic operation STILL does not work.  Dispatchers have the options of setting a individual train for automatic signal lining or they can set selected track segments to have automatic signal lining.  Additionally individual control points can have multiple moves by multiple trains 'stacked' through the control point - this includes lining switches and signals for each train in the 'stack'.  Individual signals may configured to line themselves up to 9 consecutive times for repetitive movements by that signal.  The Dispatcher that is 'on top of his game' has the ability to configure the switches and signals on his territory so that it will properly line switches and signals for all his trains for a period of several hours - and then something happens the fouls up the best laid plan.

[tree68]

Semper Vaporo

One thing that trains have always been ahead of the technological curve is in automatic steering!

And sometimes that doesn't work, either...

[Semper Vaporo]

One thing that trains have always been ahead of the technological curve is in automatic steering!

[jeffhergert]

BaltACD

 

 

jeffhergert

EMS handles some trains better than others.  Empty unit trains usually aren't a problem.  Loaded unit trains (except last night) are about second, small manifests usually aren't too bad either.  And EMS does ok through most territories.  It's where you have a series of close or sever hog backs or sags where it's train handling can become less than desirable.

Jeff 

 

 

Do the EMS systems take into account where the DPU is located in the 15K foot trains and where the hog backs are in relation to the DPU.

I get the feeling that the longer the train and the more varied the gradients encountered the more deficient the EMS software actions are.

Have YOU seen or heard of any of the EMS system designers/programmers riding trains to observe how well or poorly their programming is actually woking in the field?  I have the sneeking feeling that their algorithms were designed and encoded before PSR sized trains became the norm.  Software anticipating 9-10K foot trains would always be deficient with a 15K foot or larger train has to be hancled.  Despite being able to dial in the actual length of the train, is that length figure actually being acted upon in view of the terrain condition that you are operating upon?

 

The EMS auto throttle runs the DP separately then the head end.  I was surprised when EMS placed the DP in brake 8.  The rest of the trip when it should've used some dynamics on the rear, it wouldn't.  It let the head consist do all the dynamic braking.

Once I had some UP people and a NYAB person on a train.  It was testing the LEADER integrated into PTC.  (We've had our EMS interated with PTC for a year or so now.)  It was a manifest, but I don't recall it being one of the land barges.  More of a 'normal' sized one.  It handled the train OK for the most part.  A little bit more slack action than a human would get, but not to severe.  I think for the most part they weren't aware that in reality, we don't run until we get to a red signal.  I had to explain a few times why I went into manual mode and was going to stop to stay off crossings, etc.

A coworker who worked with this kind of tech when he was in the Air Force had an EMS tech ride with him.  (This tech originally worked with either GE or NYAB, but now worked for the railroad.)  He got him to admit that they sold the railroad that EMS could do better than 60% of their human engineers in both fuel savings and train handling.  He also admitted that the EMS wasn't living up to that.

But when you've spent millions of dollars on something, it's all a success even if it isn't.  And EMS isn't the first time that's happened on the railroad.

I've been told the Trip Optimizer algorithms are the same ones used for PTC to calculate braking distances.  

Jeff

[jeffhergert]

Lithonia Operator

Jeff, if given the choice, would most engineers draw a bye on these systems entirely?

 

Yes.  No one likes them.  Even most of the field managers, off the record, know their limitations and don't care for them.

We used to have to report to the help desk when the ESM didn't work.  We would get a ticket number for the issue.  Upon tie-up, we would also note on a specific computer form the issue and ticket number.  (Now we only report the problem, we no longer get a ticket number or have to talk to someone.)  It was said they were gathering evidence to sue one of the EMS companies because they weren't getting the benefits that company led them to believe they would receive.  I don't think they ever did that.

Jeff

[PJS1]

Tree and Jeff,

Thanks.  Very helpful insight.

[MMLDelete]

Jeff, if given the choice, would most engineers draw a bye on these systems entirely?

[BaltACD]

jeffhergert

EMS handles some trains better than others.  Empty unit trains usually aren't a problem.  Loaded unit trains (except last night) are about second, small manifests usually aren't too bad either.  And EMS does ok through most territories.  It's where you have a series of close or sever hog backs or sags where it's train handling can become less than desirable.

Jeff 

Do the EMS systems take into account where the DPU is located in the 15K foot trains and where the hog backs are in relation to the DPU.

I get the feeling that the longer the train and the more varied the gradients encountered the more deficient the EMS software actions are.

Have YOU seen or heard of any of the EMS system designers/programmers riding trains to observe how well or poorly their programming is actually woking in the field?  I have the sneeking feeling that their algorithms were designed and encoded before PSR sized trains became the norm.  Software anticipating 9-10K foot trains would always be deficient with a 15K foot or larger train has to be hancled.  Despite being able to dial in the actual length of the train, is that length figure actually being acted upon in view of the terrain condition that you are operating upon?

[jeffhergert]

tree68

 

 

PJS1

To what extent has computer technology taken over some of the tasks of operating a modern railroad locomotive? 

 

 

Both GE and NY Air Brake have systems that come close with regard to train handling.  

Just like your car, most of the functions are now computer based.

 

I turned off Trip Optimizer (GE) last night going through a series of hog backs.  (It had the DP in dynamic brake 8 when our guide lines for the territory call for the DP to be in throttle 1 or 2, pushing the slack in.  Not pulling it out.)  LEADER (NYAB) isn't much better.  They are thinking of making that official, turning off the EMS (energy managment systems) in 'challenging locations' because they have been tearing up the larger trains we are running.  (Normally, I wouldn't have turned it off.  They want it to be used as much as possible and 9 times out of 10 the ems gets away with bad train handling.  So I usually give them what they want.  Last night I just wanted to get in without having to be delayed by putting a train back together.)

I think it's because all they (ems) take into account is speed.  They don't take into account the terrain they are traversing.  In this, Trip Optimizer is worse because it 'plans' the trip and then runs the train to that plan.  If it plans on going over the top of a hill at 27 mph, it will go over the top at 27 mph.  Even if it has to back off the throttle or even go into dynamics to do it. 

LEADER at leasts recalculates more often, but even that sometimes isn't the best. I've noticed LEADER seems to compare with where it is and where it will be.  It's like it will be running at 45 mph in throttle 8 and 'see' that because of the change in grade it would be doing 55 mph in 4 or 5 miles, so it backs off the throttle now.  It loses some speed, then senses it's too slow for current conditions, so it notches up on the throttle.

The worst for both systems is going down hill.  Both at times have a bad habit of going from power to dynamics, push in the slack, then decide it needs to go back into power.  So it gets out of dynamics and lets the head end roll out.  They will do this numerous times, back and forth, on some grades.  I, and I've since noticed a few others, call this 'playing the train like an accordian.'   If a human did this and broke the train in two (or more parts), they would get censured.  The EMS doing this is called a 'mechanical failure' and is OK.

EMS handles some trains better than others.  Empty unit trains usually aren't a problem.  Loaded unit trains (except last night) are about second, small manifests usually aren't too bad either.  And EMS does ok through most territories.  It's where you have a series of close or sever hog backs or sags where it's train handling can become less than desirable.

Jeff 

[tree68]

PJS1

To what extent has computer technology taken over some of the tasks of operating a modern railroad locomotive? 

Both GE and NY Air Brake have systems that come close with regard to train handling.  

Just like your car, most of the functions are now computer based.

[PJS1]

jeffhergert

DP is an updated version of Locotrol, Locotrol lll I believe.

During communication loss (comm loss) DP consists will continue operating at the last command they received in power throttle for 90 minutes.  If it was a dynamic brake command, it will stay in dynamics until comm is restored. 

During comm loss it is possible to signal the DPs to go to idle by using the automatic air brake.  For comm loss in power, just making a service application is supposed to get the DP to idle down and cut out it's brake valve.  In dynamics during comm loss, an emergency application is needed to get the DPs to go to idle/brake valve cut out. Jeff 

Thanks for your input.

With the exception of takeoffs and landings, many of the tasks of flying a sophisticated airplane that had been handled by the pilot are now handled by the onboard computer.  The pilots put settings into the computer, i.e. desired altitude, speed, etc., and the computer actually flies the airplane.

To what extent has computer technology taken over some of the tasks of operating a modern railroad locomotive? 

[jeffhergert]

DP is an updated version of Locotrol, Locotrol lll I believe.

During communication loss (comm loss) DP consists will continue operating at the last command they received in power throttle for 90 minutes.  If it was a dynamic brake command, it will stay in dynamics until comm is restored. 

During comm loss it is possible to signal the DPs to go to idle by using the automatic air brake.  For comm loss in power, just making a service application is supposed to get the DP to idle down and cut out it's brake valve.  In dynamics during comm loss, an emergency application is needed to get the DPs to go to idle/brake valve cut out.

Jeff

[SFbrkmn]

There is a confusing forumla on exactly where mid DP's are to be placed in the train. The length permits a 20 car count either direction and a 1,000 ft barrier that cannot be outside of that. Mid Dps are a waste of time and have created much rework w/trains leaving an initial terminal out of compliance which then creates issues of the train crews being pulled from service. If there is an issue, one needs to contact their RFE or the rules hotline help desk and be goverened by their instructions. Personally I have not encountered any issues but being called for one still makes me nervous.

[Overmod]

Later report on the same thing:

https://bst-tsb.gc.ca/eng/rapports-reports/rail/1997/r97c0147/r97c0147.html

While we are on the subject of ignorance of Locotrol, what was the Canadian case involving the two midtrain helpers left configured to 'pull the other way'?  You provided the reference in a different thread but search won't find it now, and I can't tickle Google with the right search params to cough it up.

One of the most terrifying things I've seen is that initial video (the one that was taken down within 30 minutes) of the accident -- I think at Panhandle -- where the lead unit's control equipment was so rapidly and thoroughly demolished that the rear DPU kept pushing the consist at fairly high speed into the developing debris field...

[SD70Dude]

Here's a incident that resulted from improper train handling and operation of the Locotrol II system.  The current DP system has many of the same features.

https://tsb.gc.ca/eng/rapports-reports/rail/1996/r96c0086/r96c0086.html

[Overmod]

PJS1

When a long train is going through the curve, the locos on the back are out-of-sight of the lead locos.  What happens to the commands from the lead locomotive? 

One answer to this is the same as the answer to LOS in tunnels or winding ravines.  Repeaters or passive antennas.

Remember that the actual data rate being transferred across the radio is comparatively slow, so things like multipath interference are not critical, and it becomes possible even in analog to code the data for typical MU-style commands so that multiple repeating sources won't functionally interfere with each other.

A simple repeater setup might take the signals from the 'master', transmit them up the 'mountain' in the example, and then rebroadcast so that you had the equivalent of 50 microvolt contour at all points along the track (or more precisely the path of the antennas on the slaves).  The number of locations needed for this would be determined by the length of consist, number of slave locations, and a few other relatively simple factors.

It is possible that a tuned passive antenna could be provided along the track route along the 'trouble' areas, with the signal being repeater-amplified if desired.  A similar approach was used to put FM radio into the bores of the Lincoln Tunnel years ago, which is technically far more demanding than servocontrol of Woodward governors and the like.

Modern developments for the cellular-telephone industry have of course made it not only 'trivial' but cost-effective to implement remote signal assurance for any number of midtrain slaves, and to allow a computer on the head end to calculate all the different positions for throttle and DB relative to GPS/GIS to control things like node position and predictive throttle up/down for sensible train handling of very long consists over comparatively difficult profiles with minimum difficulty for the engineer (compared to what would be required to control slack with only common throttle and one-pipe air brake)

[PJS1]

Falcon48

 One interesting thing about DPU, since it is a radio based control sysytem,  is how it deals with communication failures between the head end and the DPU units, which can easily happen on a rail line.  My understanding of how this works (subject to correction by those who know more about the technology than I do) is that, when there is a communication failure, the DPU units will, at least initially, continue to operate based on their last command.  But this will stop if something happens that is out of "correspondence" with the last command. 

I know of a horseshoe curve in central Texas that is unlike the Horseshoe Curve near Altoona.  It does not have a valley between the legs of the curve.  The curve winds its way around a sizeable mountain.   Actually, it is more like a large hill, but we Texans have a habit of exaggerating stuff. 

 

When a long train is going through the curve, the locos on the back are out-of-sight of the lead locos.  What happens to the commands from the lead locomotive? 

[Falcon48]

One interesting thing about DPU, since it is a radio based control sysytem,  is how it deals with communication failures between the head end and the DPU units, which can easily happen on a rail line.  My understanding of how this works (subject to correction by those who know more about the technology than I do) is that, when there is a communication failure, the DPU units will, at least initially, continue to operate based on their last command.  But this will stop if something happens that is out of "correspondence" with the last command.

For example, if the last command was "run 8" and there is a communication failure, the DPU units will continue in run 8.  But, if there's a brake application from the head end, the computers on the DPU units will sense the drop in train line pressure that initiates a brake application, recognize it as being out of "correspondence" with the prior run 8 command, and drop the DPU units to idle without any command from the head end. Someone with more expertise than I have may want to further comment on this. 

[Rambo2]

Even at my age still learned something new today